CN113676217B - Digital predistortion correction method and device for frequency hopping communication - Google Patents

Digital predistortion correction method and device for frequency hopping communication Download PDF

Info

Publication number
CN113676217B
CN113676217B CN202110989246.6A CN202110989246A CN113676217B CN 113676217 B CN113676217 B CN 113676217B CN 202110989246 A CN202110989246 A CN 202110989246A CN 113676217 B CN113676217 B CN 113676217B
Authority
CN
China
Prior art keywords
frequency
frequency hopping
coefficient
signal
dpd
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110989246.6A
Other languages
Chinese (zh)
Other versions
CN113676217A (en
Inventor
刘颖
夏翔杰
潘文生
徐强
邵士海
马万治
沈莹
唐友喜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202110989246.6A priority Critical patent/CN113676217B/en
Publication of CN113676217A publication Critical patent/CN113676217A/en
Application granted granted Critical
Publication of CN113676217B publication Critical patent/CN113676217B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B1/00Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
    • H04B1/69Spread spectrum techniques
    • H04B1/713Spread spectrum techniques using frequency hopping
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B1/00Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
    • H04B1/02Transmitters
    • H04B1/04Circuits
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B17/00Monitoring; Testing
    • H04B17/10Monitoring; Testing of transmitters
    • H04B17/11Monitoring; Testing of transmitters for calibration
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B17/00Monitoring; Testing
    • H04B17/10Monitoring; Testing of transmitters
    • H04B17/15Performance testing
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B1/00Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
    • H04B1/02Transmitters
    • H04B1/04Circuits
    • H04B2001/0408Circuits with power amplifiers
    • H04B2001/0425Circuits with power amplifiers with linearisation using predistortion
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00Reducing energy consumption in communication networks
    • Y02D30/70Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Amplifiers (AREA)

Abstract

The invention discloses a digital predistortion correction method and a device for frequency hopping communication, wherein the method comprises the following steps: s1, according to the working frequency range f of the frequency hopping system 1 ~f 2 The working frequency range is divided into K different frequency bands, and each frequency band adopts a group of digital predistortion coefficients: s2, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,…,c K Performing initialized coefficient extraction; s3, after the extraction of the initialization coefficient is completed, carrying out digital pre-distortion processing on the frequency hopping baseband signal; s4, after carrying out digital pre-distortion processing on the frequency hopping baseband signal, sending the frequency hopping baseband signal into a power amplifier for amplification, detecting the frequency spectrum of an output signal of the power amplifier, and entering a step S5 when the digital pre-distortion performance is deteriorated and exceeds a set index requirement; s5, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,…,c K And (4) updating. The group number of the predistortion coefficients adopted by the invention can be adjusted to be less than the frequency hopping frequency point number, thereby effectively reducing the complexity of digital predistortion correction.

Description

Digital predistortion correction method and device for frequency hopping communication
Technical Field
The present invention relates to digital predistortion technologies, and in particular, to a digital predistortion correction method and apparatus for frequency hopping communications.
Background
The frequency hopping system has good anti-interference, anti-multipath and anti-fading characteristics. Nonlinear devices, such as power amplifiers, at the radio frequency front end of the frequency hopping system introduce nonlinear distortion, which causes adjacent channel interference and deterioration of the quality of the communication of the channel. Therefore, the Digital Predistortion (DPD) technique operating in the digital baseband is usually adopted to suppress the nonlinear distortion of the power amplifier. However, in the frequency hopping communication system, the power amplifier works at different frequency hopping points, and the corresponding nonlinear distortion characteristics are also different. One frequency hopping point needs to extract a group of predistortion coefficients, and multiple frequency hopping points also need to extract multiple groups of predistortion coefficients. When the frequency hopping points of the frequency hopping system increase, the number of the digital predistortion coefficient groups to be extracted also increases correspondingly, which can improve the complexity of the system.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a digital predistortion correction method and a digital predistortion correction device for frequency hopping communication.
The purpose of the invention is realized by the following technical scheme: a digital predistortion correction method for frequency hopping communications, comprising the steps of:
s1, according to the working frequency range f of a frequency hopping system 1 ~f 2 The working frequency range is divided into K different frequency bands, and each frequency band adopts a group of digital predistortion coefficients:
let the frequency range occupied by the kth frequency band be f k-1 ~f k The corresponding digital predistortion coefficient vector is c k K =1,2, \ 8230, K, K frequency bands correspond to K sets of coefficient vectors c 1 ,c 2 ,…,c K
S2, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,…,c K Performing initialized coefficient extraction;
s3, after the extraction of the initialization coefficient is completed, carrying out digital pre-distortion processing on the frequency hopping baseband signal;
s4, after carrying out digital predistortion processing on the frequency hopping baseband signal, sending the frequency hopping baseband signal into a power amplifier for amplification, detecting the frequency spectrum of an output signal of the power amplifier, and entering a step S5 when the digital predistortion performance is deteriorated and exceeds a set index requirement;
s5, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,…,c K And (4) updating.
Further, the manner of dividing the frequency in step S1 includes:
mode one, the working frequency range f 1 ~f 2 Performing frequency division of K equal parts, namely:
frequency range of frequency band 1 is
Figure BDA0003231925290000011
Frequency band 2 has a frequency range of
Figure BDA0003231925290000012
The frequency range of the frequency band k is
Figure BDA0003231925290000013
Frequency range of frequency band K
Figure BDA0003231925290000014
Figure BDA0003231925290000015
And a second mode, dividing according to the degree of frequency response change of the power amplifier:
will operate in a frequency range f 1 ~f 2 And dividing the frequency bands into K continuous frequency bands, and ensuring that the amplitude response difference in the frequency response of the power amplifier in each frequency band is less than BdB.
Further, the step S2 includes:
s201, for the kth group coefficient c k The initialization extraction:
a1, frequency band range f corresponding to the kth group of coefficients k-1 ~f k Constructing a bandwidth of f k -f k-1 Is denoted as x = [ x (1), x (2),. ·, x (N) ], and] T
a2, the bandwidth is f k -f k-1 Up-converting the broadband baseband signal x to a carrier frequency of
Figure BDA0003231925290000021
The frequency range of the up-converted RF signal is f k-1 ~f k
A3, setting the frequency range as f k-1 ~f k The radio frequency signal of (2) is amplified and excited k-1 ~f k Non-linearity within.
A4, the output signal of the power amplifier is down-converted to the baseband through the coupler and then the baseband signal y output by the power amplifier is collected k =[y k (1),y k (2),...,y k (N)] T
A5, according to the baseband power amplifier input data x and the collected kth group of baseband power amplifier output data y k Extracting the k-th group of DPD coefficients c by adopting LS algorithm k (ii) a The DPD coefficient is a digital predistortion coefficient, and in order to improve robustness, the LS algorithm is regularized:
Figure BDA0003231925290000022
where I is the identity matrix, γ is the ridge regression coefficient, Y k To adopt y k The data in (1) is a data matrix constructed according to the basis functions of the DPD model: if there are L DPD coefficients, i.e. c k =[c k1 ,c k2 ,…,c kL ] T Then Y is k Contains L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,…,y kL ](ii) a Each column vector corresponds to a basis function: the l column vector y kl Using the DPD model as shown below,
Figure BDA0003231925290000023
where x (n) is the model input, z (n) is the output, u l (x (n)) represents a basis function, typically u l (x(n))=x(n-m)|x(n-m)| p Where p is the non-linear order and m is the memory depth, in which case y kl =[u l (y k (1)),u l (y k (2)),…,u l (y k (N))] T
S202, for each frequency band, repeating the step S201 to extract the DPD coefficients to obtain K groups of DPD coefficients c 1 ,c 2 ,…,c K
Further, the step S3 includes:
s301, judging the frequency range of the current frequency hopping frequency point; if the current frequency hopping frequency point falls on f k-1 ~f k Within the frequency range of (2), the current frequency hopping frequency band is in the kth frequency band;
s302, adopting a k group of DPD coefficients c k =[c k1 ,c k2 ,…,c kL ] T Carrying out digital predistortion on a frequency hopping baseband signal corresponding to a current frequency hopping point; let v (n) be the frequency hopping baseband signal, then use c k The pre-distorted signal is calculated by:
Figure BDA0003231925290000031
and S303, up-converting the pre-distorted signal z (n) to the position of the current frequency hopping frequency point, and sending the signal to a power amplifier for amplification.
Further, in the step S5, K groups of DPD coefficients c are processed 1 ,c 2 ,…,c K The updating method comprises the following two methods:
the first method comprises the following steps: k group coefficient c k =[c k1 ,c k2 ,…,c kL ] T The update process is as follows:
b1, generating bandwidth of f k -f k-1 Wideband baseband signal x = [ x (1), x (2),.., x (N) ], x (N)] T
B2, obtaining a signal after predistortion by the wide-band baseband signal x through DPD, namely
Figure BDA0003231925290000032
Wherein c is kl (l=1,2,…,L)Is a DPD coefficient that has not been updated. Let the predistortion signal z (N) be z = [ z (1), z (2),.. Multidot.z (N) ]] T
B3, up-converting z to carrier frequency of
Figure BDA0003231925290000033
The radio frequency band of (a);
b4, enabling the radio-frequency signal after the z up-conversion to pass through a power amplifier, enabling the output signal of the power amplifier to pass through a coupler, then performing down-conversion to a baseband, and then acquiring a baseband signal y output by the power amplifier k =[y k (1),y k (2),...,y k (N)] T
B5, updating the k-th group of DPD coefficients c by ridge regression with the following formula k
Figure BDA0003231925290000034
Wherein Y is k Is a data matrix containing L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,…,y kL ]The ith column vector y kl Is y kl =[u l (y k (1)),u l (y k (2)),…,u l (y k (N))] T ,y k (1),y k (2),...,y k (N) newly acquired data corresponding to the z power amplifier output data;
b6, repeating the steps, and carrying out treatment on all K groups of DPD coefficients c 1 ,c 2 ,…,c K Updating is carried out;
and the second method comprises the following steps: gradually updating the DPD coefficient according to the frequency hopping signal, comprising the following steps:
c1, judging the frequency range of the current frequency hopping frequency point, if the current frequency hopping frequency band is in the kth frequency band, determining the kth group of DPD coefficients C k =[c k1 ,c k2 ,…,c kL ] T Updating is carried out;
c2, firstly adopting the k-th group of DPD coefficients C which are not updated k =[c k1 ,c k2 ,…,c kL ] T Digital pre-loss of frequency hopping baseband signalsIf it is, let v (n) denote the frequency hopping baseband signal as c k The pre-distorted signal is calculated by:
Figure BDA0003231925290000035
wherein c is kl (L =1,2, \8230;, L) is a DPD coefficient that has not been updated. Let the predistortion signal z (N) be z = [ z (1), z (2),.. Multidot.z (N) ]] T
C3, up-converting the pre-distorted signal z (n) to the position of the current frequency hopping frequency point, sending the signal to a power amplifier, passing the output signal of the power amplifier through a coupler, down-converting the signal to a baseband, and collecting the baseband signal t output by the power amplifier k =[y k (1),y k (2),...,y k (N)] T
C4, a coefficient C for the kth group using the following formula k And (3) updating:
Figure BDA0003231925290000041
wherein
Figure BDA0003231925290000042
Is the k-th group coefficient after updating, mu represents the updating step length, and the value of mu is more than or equal to 0 and less than or equal to 1; y is k Is a data matrix containing L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,…,y kL ]The ith column vector y kl Is y kl =[u l (y k (1)),u l (y k (2)),…,u l (y k (N))] T Wherein y is k (1),y k (2),...,y k And (N) newly acquired power amplifier output baseband data.
C5, through the steps, after the positions of the frequency hopping points exist in all K frequency bands, all K groups of DPD coefficients C can be processed 1 ,c 2 ,…,c K And (6) updating.
A digital predistortion correction device for frequency hopping communications, comprising:
the frequency band division module is used for dividing the working frequency range into K different frequency bands according to the working frequency range of the frequency hopping system, and each frequency band adopts a group of digital predistortion coefficients:
the initialization coefficient extraction module is used for extracting coefficients for initializing K groups of predistortion coefficient vectors;
the digital pre-distortion processing module is used for performing digital pre-distortion processing on the frequency hopping baseband signal after the extraction of the initialization coefficient is finished;
the performance detection module is used for performing performance detection on a frequency hopping system containing the DPD after performing digital pre-distortion processing on the frequency hopping baseband signal, detecting the frequency spectrum of the output signal of the power amplifier, and informing the coefficient updating module when the DPD performance is deteriorated and exceeds the index requirement;
a coefficient updating module for updating K groups of DPD coefficients c 1 ,c 2 ,…,c K And (6) updating.
The beneficial effects of the invention are: the invention excites the nonlinearity of the power amplifier in a broadband frequency range by designing a broadband signal and corrects the nonlinearity in the broadband frequency range by using a group of DPD coefficients. Therefore, as long as the frequency hopping frequency point is in the broadband frequency range, the extracted group of coefficients can be adopted for DPD correction. For a frequency hopping system with a plurality of frequency hopping points, the invention only needs to extract K groups of DPD coefficients. Where K can be adjusted according to complexity. K can be set relatively small to ensure low complexity. The method adopts a ridge regression method to calculate the DPD coefficient, and can improve the robustness of DPD coefficient extraction. The extracted DPD coefficient can be better adapted to different frequency hopping frequency points in the corresponding frequency range. The method comprises the initialization calculation process of the DPD coefficient and the updating process of the DPD coefficient, is beneficial to continuously optimizing the predistortion coefficient and improving the accuracy of predistortion processing.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a schematic illustration of frequency division in an embodiment;
fig. 3 is a schematic diagram of the device of the present invention.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
As shown in fig. 1, a digital predistortion correction method for frequency hopping communication includes the following steps:
s1, according to the working frequency range f of the frequency hopping system 1 ~f 2 The working frequency range is divided into K different frequency bands, and each frequency band adopts a group of digital predistortion coefficients: if the frequency point falls within the frequency band, the corresponding DPD coefficient of the frequency band is used for correction, and the frequency band division is as shown in the following FIG. 2
Let the frequency range occupied by the kth frequency band be f k-1 ~f k The corresponding digital predistortion coefficient vector is c k K =1,2, \ 8230, K, K frequency bands correspond to K sets of coefficient vectors c 1 ,c 2 ,…,c K
The frequency division mode comprises the following steps:
mode one, the working frequency range f 1 ~f 2 Performing frequency division of K equal parts, namely:
frequency band 1 has a frequency range of
Figure BDA0003231925290000051
Frequency band 2 has a frequency range of
Figure BDA0003231925290000052
The frequency range of the frequency band k is
Figure BDA0003231925290000053
Frequency range of frequency band K
Figure BDA0003231925290000054
Figure BDA0003231925290000055
For example, the operating frequency range of a frequency hopping systemEnclose as f 1 =1900MHz to f 2 And =2100MHz, i.e. all frequency hopping bins are in the range of 1900MHz to 2100 MHz. At this time, the frequency hopping band can be divided into K =4 equally divided bands of 1900MHz to 1950MHz,1950MHz to 2000MHz,2000MHz to 2050MHz, and 2050MHz to 2010MHz, respectively.
And a second mode, dividing according to the degree of frequency response change of the power amplifier:
will operate in a frequency range f 1 ~f 2 And dividing the frequency bands into K continuous frequency bands, and ensuring that the amplitude response difference in the frequency response of the power amplifier in each frequency band is less than BdB. The BdB can be adjusted according to actual needs, and typical BdB can be selected to be 0.5dB and 1dB.
S2, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,…,c K Performing initialized coefficient extraction;
the step S2 includes:
s201, for the kth group coefficient c k The initialization extraction:
a1, frequency band range f corresponding to the k group of coefficients k-1 ~f k Constructing a bandwidth of f k -f k-1 Is denoted as x = [ x (1), x (2),. ·, x (N) ], and] T
for example, when the k =1 frequency band ranges from 1900MHz to 1950MHz, a broadband signal with a bandwidth of 50MHz is constructed.
A2, the bandwidth is f k -f k-1 Up-converting the broadband baseband signal x to a carrier frequency of
Figure BDA0003231925290000061
The frequency range of the up-converted RF signal is f k-1 ~f k
A3, setting the frequency range as f k-1 ~f k The radio frequency signal of (2) is amplified and excited k-1 ~f k Internal non-linearity.
A4, the output signal of the power amplifier passes through the coupler, then down-conversion is carried out to the baseband, and the baseband output by the power amplifier is collectedSignal y k =[y k (1),y k (2),...,y k (N)] T
A5, according to the baseband power amplifier input data x and the k-th group baseband power amplifier output data y k Extracting the k-th group of DPD coefficients c by adopting LS algorithm k (ii) a The DPD coefficient is a digital predistortion coefficient, and in order to improve robustness, the LS algorithm is regularized:
Figure BDA0003231925290000062
where I is the identity matrix and γ is the ridge regression coefficient, (γ may also be referred to as the regularization factor). Typical values of γ may be selected from 0.1,0.01,0.001,0.0001, 0.00001; y is k To adopt y k The data in (1) is a data matrix constructed according to the basis functions of the DPD model: if there are L DPD coefficients, i.e. c k =[c k1 ,c k2 ,…,c kL ] T Then Y is k Contains L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,…,y kL ](ii) a Each column vector corresponds to a basis function: the l column vector y kl Using the DPD model as shown below,
Figure BDA0003231925290000063
where x (n) is the model input, z (n) is the output, u l (x (n)) represents a basis function, typically u l (x(n))=x(n-m)|x(n-m)| p Where p is the nonlinear order and m is the memory depth, then y kl =[u l (y k (1)),u l (y k (2)),…,u l (y k (N))] T
S202, for each frequency band, repeating the step S201 to extract the DPD coefficients to obtain K groups of DPD coefficients c 1 ,c 2 ,…,c K
S3, after the extraction of the initialization coefficient is completed, carrying out digital pre-distortion processing on the frequency hopping baseband signal;
s4, after carrying out digital pre-distortion processing on the frequency hopping baseband signal, sending the frequency hopping baseband signal into a power amplifier for amplification, detecting the frequency spectrum of an output signal of the power amplifier, and entering a step S5 when the digital pre-distortion performance is deteriorated and exceeds a set index requirement;
if the digital predistortion performance does not exceed the set index requirement, the coefficient does not need to be updated, and the current coefficient vector is continuously utilized for predistortion treatment.
In the embodiment of the present application, the index generally adopts ACLR (adjacent channel leakage ratio) of the spectrum of the power amplifier output signal, and the set index requirement generally means that the ACLR (adjacent channel leakage ratio) index of the spectrum of the power amplifier output signal needs to be smaller than-45 dBc. At the moment, the frequency spectrum of the output signal of the power amplifier is detected, if the ACLR index is found to be larger than-45 dBc, the performance is considered to be deteriorated, and the index requirement is exceeded.
S5, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,…,c K And (6) updating.
Further, the step S3 includes:
s301, judging the frequency range of the current frequency hopping frequency point; if the current frequency hopping frequency point falls on f k-1 ~f k Within the frequency range of (2), the current frequency hopping band is in the kth frequency band;
s302, adopting a k group of DPD coefficients c k =[c k1 ,c k2 ,…,c kL ] T Carrying out digital predistortion on a frequency hopping baseband signal corresponding to a current frequency hopping point; let the frequency hopping baseband signal be v (n), then use c k The pre-distorted signal is calculated by:
Figure BDA0003231925290000071
and S303, up-converting the pre-distorted signal z (n) to the position of the current frequency hopping frequency point, and sending the signal to a power amplifier for amplification.
Further, in the step S5, K groups of DPD coefficients c are processed 1 ,c 2 ,…,c K The updating method comprises the following two stepsThe method comprises the following steps:
the first method comprises the following steps: coefficient of k group c k =[c k1 ,c k2 ,…,c kL ] T The update process is as follows:
b1, generating bandwidth of f k -f k-1 Is used to generate a wideband baseband signal of (1) = [ x (1), x (2) ], x (N)] T
B2, obtaining a signal after predistortion by passing the broadband baseband signal x through DPD, namely
Figure BDA0003231925290000072
Wherein c is kl (L =1,2, \8230;, L) is the DPD coefficient that has not been updated. Let the predistortion signal z (N) be z = [ z (1), z (2),.. Multidot.z (N) ]] T
B3, up-converting z to a carrier frequency of
Figure BDA0003231925290000073
The radio frequency band of (c);
b4, enabling the radio-frequency signal after the z up-conversion to pass through a power amplifier, enabling the output signal of the power amplifier to pass through a coupler, then performing down-conversion to a baseband, and then acquiring a baseband signal y output by the power amplifier k =[y k (1),y k (2),...,y k (N)] T
B5, updating the kth group of DPD coefficients c by ridge regression of the following formula k
Figure BDA0003231925290000074
Wherein Y is k Is a data matrix containing L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,…,y kL ]The l-th column vector y kl Is y kl =[u l (y k (1)),u l (y k (2)),…,u l (y k (N))] T ,y k (1),y k (2),...,y k (N) newly acquired data corresponding to the z power amplifier output data;
b6, repeating the steps, and carrying out treatment on all K groups of DPD coefficients c 1 ,c 2 ,…,c K Updating is carried out;
and the second method comprises the following steps: gradually updating the DPD coefficient according to the frequency hopping signal, comprising the following steps:
c1, judging the frequency range of the current frequency hopping frequency point, if the current frequency hopping frequency band is in the kth frequency band, determining the kth group of DPD coefficients C k =[c k1 ,c k2 ,…,c kL ] T Updating is carried out;
c2, firstly adopting the k-th group of DPD coefficients C which are not updated k =[c k1 ,c k2 ,…,c kL ] T Performing digital predistortion on the frequency hopping baseband signal, and using c to mark the frequency hopping baseband signal as v (n) k The pre-distorted signal is calculated by:
Figure BDA0003231925290000081
wherein c is kl (L =1,2, \8230;, L) is a DPD coefficient that has not been updated. Let the predistortion signal z (N) be z = [ z (1), z (2),.. Multidot.z (N) ]] T
C3, up-converting the pre-distorted signal z (n) to the position of the current frequency hopping frequency point, sending the signal to a power amplifier, passing the output signal of the power amplifier through a coupler, down-converting the signal to a baseband, and collecting the baseband signal y output by the power amplifier k =[y k (1),y k (2),...,y k (N)] T
C4, a coefficient C for the kth group using the following formula k Updating:
Figure BDA0003231925290000082
wherein
Figure BDA0003231925290000083
Is the k-th group coefficient after updating, mu represents the updating step length, and the value of mu is more than or equal to 0 and less than or equal to 1; y is k Is a data matrix comprising L columnsVector of Y k =[y k1 ,y k2 ,y k3 ,…,y kL ]The ith column vector y kl Is y kl =[u l (y k (1)),u l (y k (2)),…,u l (y k (N))] T Wherein y is k (1),y k (2),...,y k And (N) newly acquired power amplifier output baseband data.
C5, through the steps, after the positions of the frequency hopping points exist in all K frequency bands, all K groups of DPD coefficients C can be processed 1 ,c 2 ,…,c K And (6) updating.
As shown in fig. 3, a digital predistortion correction apparatus for frequency hopping communication, includes:
the frequency band division module is used for dividing the working frequency range into K different frequency bands according to the working frequency range of the frequency hopping system, and each frequency band adopts a group of digital predistortion coefficients:
the initialization coefficient extraction module is used for extracting the coefficient for initializing the K groups of predistortion coefficient vectors;
the digital pre-distortion processing module is used for performing digital pre-distortion processing on the frequency hopping baseband signal after the extraction of the initialization coefficient is finished;
the performance detection module is used for performing performance detection on a frequency hopping system containing DPD after digital pre-distortion processing is performed on a frequency hopping baseband signal, detecting the frequency spectrum of a power amplifier output signal, and informing the coefficient updating module when the DPD performance is deteriorated and exceeds index requirements;
a coefficient updating module for updating K groups of DPD coefficients c 1 ,c 2 ,…,c K And (4) updating.
The above examples are only intended to illustrate the technical solution of the present invention, and not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (5)

1. A digital predistortion correction method for frequency hopping communications, characterized by: the method comprises the following steps:
s1, according to the working frequency range f of a frequency hopping system 1 ~f 2 The working frequency range is divided into K different frequency bands, and each frequency band adopts a group of digital predistortion coefficients:
let the frequency range occupied by the kth frequency band be f k-1 ~f k The corresponding digital predistortion coefficient vector is c k K =1, 2.. K, then K frequency bands correspond to K sets of coefficient vectors c 1 ,c 2 ,...,c K
S2, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,...,c K Performing initialized coefficient extraction;
the step S2 includes:
s201, for k group of coefficients c k The initialization extraction:
a1, frequency band range f corresponding to the k group of coefficients k-1 ~f k Constructing a bandwidth of f k -f k-1 Is denoted as x = [ x (1), x (2),. ], x (N)] T
A2, setting the bandwidth as f k -f k-1 Up-converting the broadband baseband signal x to a carrier frequency of
Figure FDA0003800065710000011
The frequency range of the up-converted RF signal is f k-1 ~f k
A3, setting the frequency range as f k-1 ~f k The radio frequency signal of (2) is amplified and excited k-1 ~f k Internal non-linearity;
a4, the output signal of the power amplifier is converted to a baseband through the coupler and then the baseband signal y output by the power amplifier is collected k =[y k (1),y k (2),...,y k (N)] T
A5, according to the baseband power amplifier input data x and the collected kth group of baseband power amplifier output data y k Extracting the k-th group of DPD coefficients c by adopting LS algorithm k (ii) a The DPD coefficient is a digital predistortion coefficient, and in order to improve robustness, the LS algorithm is regularized:
Figure FDA0003800065710000012
wherein I is a unit matrix, gamma is a ridge regression coefficient, and Y is k To adopt y k The data in (1) is a data matrix constructed according to the basis functions of the DPD model: if there are L DPD coefficients, i.e. c k =[c k1 ,c k2 ,...,c kL ] T Then Y is k Contains L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,...,y kL ](ii) a Each column vector corresponds to a basis function: the l column vector y kl Using the DPD model as shown below,
Figure FDA0003800065710000013
where x (n) is the model input, z (n) is the output, u l (x (n)) represents a basis function, typically u l (x(n))=x(n-m)|x(n-m)| p Where p is the nonlinear order and m is the memory depth, then y kl =[u l (y k (1)),u l (y k (2)),...,u l (y k (N))] T
S202, for each frequency band, repeating the step S201 to extract the DPD coefficients to obtain K groups of DPD coefficients c 1 ,c 2 ,...,c K
S3, after the extraction of the initialization coefficient is completed, carrying out digital pre-distortion processing on the frequency hopping baseband signal;
s4, after carrying out digital predistortion processing on the frequency hopping baseband signal, sending the frequency hopping baseband signal into a power amplifier for amplification, detecting the frequency spectrum of an output signal of the power amplifier, and entering a step S5 when the digital predistortion performance is deteriorated and exceeds a set index requirement;
s5, for K groups of digital predistortion coefficient vectors c 1 ,c 2 ,...,c K And (6) updating.
2. A digital predistortion correction method for frequency hopping communications according to claim 1, characterized in that: the frequency division mode in step S1 includes:
mode one, the working frequency range f 1 ~f 2 Performing frequency division of K equal parts, namely:
frequency range of frequency band 1 is
Figure FDA0003800065710000021
Frequency band 2 has a frequency range of
Figure FDA0003800065710000022
Frequency range of frequency band k of
Figure FDA0003800065710000023
Frequency range of frequency band K
Figure FDA0003800065710000024
Figure FDA0003800065710000025
And a second mode, dividing according to the degree of frequency response change of the power amplifier:
will operate in a frequency range f 1 ~f 2 And dividing the frequency bands into K continuous frequency bands, and ensuring that the amplitude response difference in the frequency response of the power amplifier in each frequency band is less than BdB.
3. A digital predistortion correction method for frequency hopping communications according to claim 1, characterized in that: the step S3 includes:
s301, judging the frequency range of the current frequency hopping frequency point; if the current frequency hopping frequency point falls on f k-1 ~f k Within the frequency range of (2), the current frequency hopping band is in the kth frequency band;
s302, adopting a k group of DPD coefficients c k =[c k1 ,c k2 ,...,c kL ] T Carrying out digital predistortion on a frequency hopping baseband signal corresponding to a current frequency hopping point; let the frequency hopping baseband signal be v (n), then use c k The pre-distorted signal is calculated by:
Figure FDA0003800065710000026
and S303, up-converting the pre-distorted signal z (n) to the position of the current frequency hopping frequency point, and sending the signal to a power amplifier for amplification.
4. A digital predistortion correction method for frequency hopping communications according to claim 1, characterized in that: for K groups of DPD coefficients c in the step S5 1 ,c 2 ,...,c K The updating method comprises the following two methods:
the first method comprises the following steps: k group coefficient c k =[c k1 ,c k2 ,...,c kL ] T The updating process is as follows:
b1, generating bandwidth of f k -f k-1 Is used to generate a wideband baseband signal of (1) = [ x (1), x (2) ], x (N)] T
B2, obtaining a signal after predistortion by the wide-band baseband signal x through DPD, namely
Figure FDA0003800065710000027
Wherein c is kl (L =1, 2.. Gtoreq.l) is a DPD coefficient that has not been updated, and the predistortion signal z (N) is denoted as z = [ z (1), z (2),. Gtoreq.z (N)] T
B3, up-converting z toA carrier frequency of
Figure FDA0003800065710000031
The radio frequency band of (c);
b4, the radio-frequency signal after the z up-conversion is subjected to power amplification, the output signal of the power amplification is subjected to down-conversion to a baseband through a coupler, and then the baseband signal y output by the power amplification is collected k =[y k (1),y k (2),...,y k (N)] T
B5, updating the kth group of DPD coefficients c by ridge regression of the following formula k
Figure FDA0003800065710000032
Wherein Y is k Is a data matrix containing L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,...,y kL ]The l-th column vector y kl Is y kl =[u l (y k (1)),u l (y k (2)),...,u l (y k (N))] T ,y k (1),y k (2),...,y k (N) newly acquired data corresponding to the z power amplifier output data;
b6, repeating the steps, and carrying out treatment on all K groups of DPD coefficients c 1 ,c 2 ,...,c K Updating is carried out;
and the second method comprises the following steps: gradually updating the DPD coefficient according to the frequency hopping signal, comprising the following steps:
c1, judging the frequency range of the current frequency hopping frequency point, if the current frequency hopping frequency band is in the kth frequency band, determining the kth group of DPD coefficients C k =[c k1 ,c k2 ,...,c kL ] T Updating is carried out;
c2, firstly adopting the k-th group of DPD coefficients C which are not updated k =[c k1 ,c k2 ,...,c kL ] T Digitally pre-distorting the frequency hopping baseband signal, and marking the frequency hopping baseband signal as v (n), using c k Calculating the pre-distorted signal by:
Figure FDA0003800065710000033
Wherein c is kl (L =1, 2.. So, L) is a DPD coefficient which has not been updated, and the predistortion signal z (N) is denoted as z = [ z (1), z (2),. So, z (N)] T
C3, up-converting the pre-distorted signal z (n) to the position of the current frequency hopping frequency point, sending the signal to a power amplifier, passing the output signal of the power amplifier through a coupler, down-converting the signal to a baseband, and collecting the baseband signal y output by the power amplifier k =[y k (1),y k (2),...,y k (N)] T
C4, a coefficient C of the kth group using the formula k And (3) updating:
Figure FDA0003800065710000034
wherein
Figure FDA0003800065710000035
Is the k-th group coefficient after updating, mu represents the updating step length, and the value of mu is more than or equal to 0 and less than or equal to 1; y is k Is a data matrix containing L column vectors, denoted as Y k =[y k1 ,y k2 ,y k3 ,...,y kL ]The ith column vector y kl Is y kl =[u l (y k (1)),u l (y k (2)),...,u l (y k (N))] T Wherein y is k (1),y k (2),...,y k (N) newly acquired power amplifier output baseband data;
c5, through the steps, after the positions of the frequency hopping points exist in all K frequency bands, all K groups of DPD coefficients C can be processed 1 ,c 2 ,...,c K And (6) updating.
5. A digital predistortion correction device for frequency hopping communication, based on the method of any one of claims 1 to 4, characterized in that: the method comprises the following steps:
the frequency band division module is used for dividing the working frequency range into K different frequency bands according to the working frequency range of the frequency hopping system, and each frequency band adopts a group of digital predistortion coefficients:
the initialization coefficient extraction module is used for performing initialized coefficient extraction on the K groups of predistortion coefficient vectors;
the digital pre-distortion processing module is used for performing digital pre-distortion processing on the frequency hopping baseband signal after the extraction of the initialization coefficient is finished;
the performance detection module is used for performing performance detection on a frequency hopping system containing the DPD after performing digital pre-distortion processing on the frequency hopping baseband signal, detecting the frequency spectrum of the output signal of the power amplifier, and informing the coefficient updating module when the DPD performance is deteriorated and exceeds the index requirement;
a coefficient updating module for updating K groups of DPD coefficients c 1 ,c 2 ,…,c K And (4) updating.
CN202110989246.6A 2021-08-26 2021-08-26 Digital predistortion correction method and device for frequency hopping communication Active CN113676217B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110989246.6A CN113676217B (en) 2021-08-26 2021-08-26 Digital predistortion correction method and device for frequency hopping communication

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110989246.6A CN113676217B (en) 2021-08-26 2021-08-26 Digital predistortion correction method and device for frequency hopping communication

Publications (2)

Publication Number Publication Date
CN113676217A CN113676217A (en) 2021-11-19
CN113676217B true CN113676217B (en) 2022-11-08

Family

ID=78546761

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110989246.6A Active CN113676217B (en) 2021-08-26 2021-08-26 Digital predistortion correction method and device for frequency hopping communication

Country Status (1)

Country Link
CN (1) CN113676217B (en)

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102594749A (en) * 2012-02-28 2012-07-18 中兴通讯股份有限公司 Digital pre-distortion processing method and device

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3635886A4 (en) * 2017-06-09 2020-04-15 Nanosemi, Inc. Subsampled linearization system
CN109150240B (en) * 2018-07-18 2020-07-14 中国电子科技集团公司第七研究所 Digital predistortion on-line training method for wireless frequency hopping communication system
CN111585608B (en) * 2020-05-07 2022-03-22 中国人民解放军63921部队 Adaptive digital predistortion method for broadband spread frequency hopping system

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102594749A (en) * 2012-02-28 2012-07-18 中兴通讯股份有限公司 Digital pre-distortion processing method and device

Also Published As

Publication number Publication date
CN113676217A (en) 2021-11-19

Similar Documents

Publication Publication Date Title
CN106506417B (en) Narrow-band feedback digital predistortion system and method
CN111436225B (en) Predistortion circuit of wireless transmitter and method for generating predistortion baseband signal
EP2575310B1 (en) Method and device for digital baseband predistortion
US8577311B2 (en) Predistortion of amplifier input signals
US6741662B1 (en) Transmitter linearization using fast predistortion
WO2016151518A1 (en) Method and apparatus for multiband predistortion using time-shared adaptation loop
CN106170918B (en) Linear damage is subtracted for nonlinear impairments digital pre-distortion error signal
US10892786B1 (en) Digital predistortion with out-of-band and peak expansion regularization
WO2013056660A1 (en) Predistortion circuit, wireless communication unit and method for coefficient estimation
CN113037226B (en) Digital predistortion design method and device based on adaptive step length clipping method
CN102281221A (en) Nonlinear systematic distortion correction device and method
CN107276546A (en) A kind of digital pre-distortion processing method and device
CN101610093B (en) Predistortion device, predistortion system and predistortion method
CN102611661B (en) Predistortion device and method based on precise inverse solution memory polynomial model equation
CN112019221B (en) Signal processing method, device and storage medium
CN103929136B (en) Broadband mobile communication LTE system transmitter power amplifier predistortion processing method
US9787335B1 (en) Method and apparatus for determining validity of samples for digital pre-distortion apparatus
Liu et al. On the robustness of look-up table digital predistortion in the presence of loop delay error
CN102055696A (en) Digital predistortion system for inhibiting noise of feedback signal
WO2015188578A1 (en) Nonlinear system distortion correction device and method
CN113676217B (en) Digital predistortion correction method and device for frequency hopping communication
CN116436539B (en) Nonlinear calibration device and method for amplifier
CN102983820B (en) Nonlinear injection type linearization system and digital predistortion method
CN116232345A (en) Predistortion system with target spectrum emission for wireless communication
CN206922797U (en) A kind of short range digital pre-distortion device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant